About Me

I engineer intelligent systems that bridge algorithmic potential and real-world reliability. My expertise lies in architecting and deploying full-stack AI solutions from embedded edge devices to noise-resistant NLP models that perform under the complex and uncertain conditions of actual use.

My work is guided by a philosophy of pragmatic innovation: a commitment to solving tangible problems with robust, efficient, and ethically considered technology. I am not just interested in what a model can achieve on a benchmark, but in how it functions in production, how it handles messy, subjective data, and how it respects user privacy and system constraints.

Core Capabilities

• ML/NLP: Designing and fine-tuning state-of-the-art models (e.g., Transformer architectures like DeBERTa) with innovative training paradigms to overcome real-world data challenges like significant label noise and low inter-annotator agreement.

• Edge AI & Embedded Systems: Developing privacy-first, low-latency systems by leveraging edge computing, TensorFlow Lite, and real-time C/C++ on microcontrollers and embedded Linux platforms.

• End-to-End System Architecture: Building and integrating complex, decoupled systems—such as virtualized environments and custom networking pipelines—to ensure scalable and secure deployment.

Select Project Highlights

• Robust Automated Assessment: Achieved a state-of-the-art Quadratic Weighted Kappa (QWK of 0.72) on a highly subjective essay-grading task by implementing a novel co-teaching paradigm that reduced the impact of label noise by 30%.

• Privacy-Preserving Accessibility Tech: Engineered an edge-AI assistant for the visually impaired, achieving sub-100ms inference latency via an optimized TensorFlow Lite pipeline on an ARM64 edge device, ensuring all data processing occurs locally.

• High-Performance Embedded Control: Sustained a 125 Hz control loop for an autonomous robot by implementing a discrete-time PID controller and adaptive sensor calibration, enabling reliable navigation at 0.4 m/s across varying terrain.

I am actively seeking a master degree to apply my cross-disciplinary skills in Machine Learning Engineering, Applied Research, or Embedded Systems to ambitious, impactful problems. I am particularly interested in roles that challenge me to push the boundaries of deploying AI outside the lab.

Let's connect to discuss how my approach to building reliable, real-world intelligent systems can bring value to your team or lab.